| Summary: | 碩士 === 國立成功大學 === 電機工程學系 === 102 === Gene similarity is helpful for predicting various biological mechanisms. With the advance of biotechnologies, various types of biological data, such as sequence similarity, functional similarity, co-expression, co-protein interaction, and co-regulation are available for studying gene relations. These different types of biological data are figured by types of features and mutually affected by each other.
Statistically quantifying the linkage among different types of biological data is helpful to discover the linkages between features and select proper feature to alternate desired biological feature by inadequate data.
In the previous works, the linkage between different types of biological data is implicitly applied to validate or quantify an unknown biological data by other available data. However the linkages between different types of biological data are simply mentioned in these works without a systematic analysis.
The present work gives a comprehensive study of linkage among co-expression, co-regulation, co-function, protein-protein interaction similarity and sequence similarity. The five types of biological data were selected because (i) they are widely available in many species (availability) and (ii) there are existing gene/protein relation measures based on them (popularity).
The linkages between features are analyzed by mean-value curve, and further identified into entirely implicated, partially implicated, and obscure. Base on the identified linkages, proper predictors of each type of data are presented. Finally, these linkages among features are globally analyzed and some biological mechanisms are revealed from these globally-analyzed relations between types of data.
|
|---|
